Company
Date Published
Author
Reinhard Weber
Word count
2319
Language
American English
Hacker News points
None

Summary

In the fourth installment of a blog series on managing Dynatrace Managed at scale, the author explores optimizing anomaly detection settings using Dynatrace's raw problem and event data. The author highlights the challenges of managing problem noise across thousands of environments and introduces strategies to reduce the number of support tickets by visualizing historical problem data with a "Swimlane Visualization" and conducting statistical analyses. The author defines an "unhealthy" environment by the presence of at least one open problem and discusses the benefits of consolidating overlapping problems into singular "unhealthy situations" to streamline ITSM processes. By adjusting notification times and tweaking anomaly detection parameters, the author successfully reduces alert noise and unnecessary notifications, particularly for short-duration problems. The process involves iteratively analyzing data and refining settings, which ultimately leads to a 50% reduction in detected problems without loss of visibility. The blog concludes with a teaser for the next part, which will address Dynatrace configuration as code, emphasizing its relevance for both large-scale and smaller setups.